Getting Started
The following example demonstrates how to use the GenAI-powered Document Insights functionality to summarize a Word document and ask questions about it:
The following code snippet is valid for Azure Open AI 9.3. The specific IChatClient initialization may be different according to the specific version.
For .NET 8+ (Target OS Windows) with Packages for .NET 8 and .NET 9 for Windows, an IEmbeddingsStorage implementation is required for the PartialContextQuestionProcessor.
[C#] Example 1: Using GenAI-powered Document Insights
public async void ProcessinputFlowDocWithAI()
{
// Load the Docx document
string filePath = @"path\to\your\document.docx";
DocxFormatProvider formatProvider = new DocxFormatProvider();
RadFlowDocument flowDocument;
using (FileStream fs = File.OpenRead(filePath))
{
flowDocument = formatProvider.Import(fs, TimeSpan.FromSeconds(10));
}
// Convert the document to a simple text representation
ISimpleTextDocument plainDoc = flowDocument.ToSimpleTextDocument(TimeSpan.FromSeconds(10));
// Set up the AI client (Azure OpenAI in this example)
string key = "AZUREOPENAI_KEY";
string endpoint = "AZUREOPENAI_ENDPOINT";
string model = "gpt-4o-mini";
Azure.AI.OpenAI.AzureOpenAIClient azureClient = new AzureOpenAIClient(
new Uri(endpoint),
new Azure.AzureKeyCredential(key),
new Azure.AI.OpenAI.AzureOpenAIClientOptions());
OpenAI.Chat.ChatClient chatClient = azureClient.GetChatClient(model);
IChatClient iChatClient = new OpenAIChatClient(chatClient);
int maxTokenCount = 128000;
// 1. Summarize the document
using (SummarizationProcessor summarizationProcessor = new SummarizationProcessor(iChatClient, maxTokenCount))
{
// Handle resources calculation event to control token usage
summarizationProcessor.SummaryResourcesCalculated += (sender, e) =>
{
Console.WriteLine($"Estimated calls required: {e.EstimatedCallsRequired}");
Console.WriteLine($"Estimated tokens required: {e.EstimatedTokensRequired}");
// Confirm if the operation should continue
e.ShouldContinueExecution = true;
};
string summary = await summarizationProcessor.Summarize(plainDoc);
Console.WriteLine("Document Summary:");
Console.WriteLine(summary);
}
// 2. Answer questions using partial context (recommended for efficiency)
#if NET8_0_WINDOWS
using (PartialContextQuestionProcessor partialContextQuestionProcessor = new PartialContextQuestionProcessor(iChatClient, maxTokenCount, plainDoc))
{
string question = "What are the main findings in the document?";
string answer = await partialContextQuestionProcessor.AnswerQuestion(question);
Console.WriteLine($"Question: {question}");
Console.WriteLine($"Answer: {answer}");
}
#else
IEmbeddingsStorage embeddingsStorage = new OllamaEmbeddingsStorage();
using (PartialContextQuestionProcessor partialContextQuestionProcessor =
new PartialContextQuestionProcessor(iChatClient, embeddingsStorage, maxTokenCount, plainDoc))
{
string question = "What are the main findings in the document?";
string answer = await partialContextQuestionProcessor.AnswerQuestion(question);
Console.WriteLine($"Question: {question}");
Console.WriteLine($"Answer: {answer}");
}
#endif
// 3. Answer questions using complete context (for smaller documents)
using (CompleteContextQuestionProcessor completeContextQuestionProcessor =
new CompleteContextQuestionProcessor(iChatClient, maxTokenCount))
{
string question = "What is the conclusion of the document?";
string answer = await completeContextQuestionProcessor.AnswerQuestion(plainDoc, question);
Console.WriteLine($"Question: {question}");
Console.WriteLine($"Answer: {answer}");
}
}
When you run this code, the AI will process your document, generate a summary, and answer your questions.